Molecular characterization and personalized approaches to non-Hodgkin lymphoma from circulating tumor DNA Biological and clinical heterogeneity between patients remain inherent barriers to improving cancer outcomes. This heterogeneity makes prediction of therapeutic outcomes and personalized treatment approaches a major challenge in clinical oncology. This is exemplified by diffuse large B-cell lymphoma (DLBCL), the most common non-Hodgkin lymphoma in adults. Over the last 30 years, significant strides have been made in unraveling the inter-patient heterogeneity of this disease, leading to identification of molecular subtypes based on the cell-of-origin and risk stratification tools based on clinical and radiographic factors. Recently, the field of genomics has made additional strides in identifying the specific genes that drive lymphomas in individual patients, which has in turn led to a revolution in personalized medicine. Nevertheless and despite these advances, current personalized approaches to therapy have failed to improve outcomes for patients with DLBCL. During the last decade, novel methods to detect cell-free tumor-derived DNA, or circulating tumor DNA (ctDNA), have emerged. We previously applied Cancer Personalized Profiling by Deep Sequencing (CAPP- Seq), a targeted sequencing approach for ctDNA detection, to patients with DLBCL. Additionally, we recently explored the response dynamics of ctDNA in DLBCL patients receiving standard therapy, defining robust molecular response criteria after as few as 21 days. Detection of ctDNA therefore provides an opportunity for both improved understanding of tumor genotype and phenotype, including response to therapy, opening the door to personalized approaches to diagnosis and disease management. In this proposal, I will further extend ctDNA detection by CAPP-Seq to develop methods to comprehensively molecularly characterize DLBCL directly from the blood plasma, including genome-wide copy-number profiling and detection of phased haplotypes, a unique entity in B-cell malignancies (Aim 1). I will then refine and validate a novel framework, called the Continuous Individualized Risk Index (CIRI), to integrate genomic information with ctDNA molecular response criteria to build a clinically useful personalized model of risk for DLBCL patients (Aim 2). Finally, I will apply these tools to study the genetics and molecular response dynamics of DLBCL patients receiving chimeric antigen receptor (CAR) T- cells, an emerging therapy for relapsed and refractory DLBCL, which remains an area of clinical need (Aim 3). This proposal will be carried out at the Stanford University School of Medicine, under the mentorship of Ash Alizadeh, MD/PhD. Through completion of this proposal, I will gain the relevant experience in computational biology and biomedical data science to successfully launch a career as an independent investigator focused on developing and translating new technologies for patients with lymphoma.